Key points are not available for this paper at this time.
Target tracking algorithms have to operate in an environment of uncertain measurement origin, in the presence of possibly non-detected target measurements as well as clutter measurements from unwanted scatterers. Integral part of expressions for data association probabilities is the estimate of clutter density. A priori knowledge of the clutter density may be derived in the form of the closed expression, whilst in the case of environmentally caused clutter; clutter measurement density has to be estimated using measurements from a number of scans, using a clutter map. This paper presents, as well as compares, three different clutter map estimators. The first is the classic clutter map estimator, which averages the number of measurements over a number of scans. This is a biased estimator of the inverse of clutter density. The second clutter map estimator uses spatial characteristics of the multi-dimensional Poisson process, and is termed spatial clutter map estimator. The third clutter map estimator uses temporal characteristics of the Poisson process, thus this estimator is termed temporal clutter map estimator. Both spatial and temporal clutter map produce an unbiased estimate of the inverse of clutter measurement density.
Mušicki et al. (Sat,) studied this question.